Welcome to the LS-DYNA Product Space

Hi everyone, I am trying to set up a DUALCESE model to simulate a gas jet but I am getting a hierchy error. From what I have learned I need to define all my DUALCESE keywords under the DUALCESE_MODEL hierchy but the DUALCESE_MODEL keyword seems to dissappear from the keyword deck after I save the project.  Can someone please guide me on ho to set-up a correct hierchy for the problem or if there are any documents availabe, would really appreaciate it. Thanks in advance.  

<p class="PDq2pG_selectionAnchorContainer" data-start="87" data-end="93"><span style="font-size: 12pt;">Hello,</span></p><p class="" data-start="95" data-end="245"><span style="font-size: 12pt;">I am using LS-DYNA 2025 R2 on Windows to obtain the unloaded configuration of a pressurised hyperelastic model using *CONTROL_REFERENCE_CONFIGURATION.</span></p><p data-start="247" data-end="293"><span style="font-size: 12pt;">The analysis is launched with the CASE option, as required:</span></p><p data-start="295" data-end="374"><span style="font-size: 12pt;">mpiexec -np %NCPU% -aa -a "%SOLVER%" CASE i=run_zero_p_config.k memory=%MEMORY%</span></p><p data-start="376" data-end="520"><span style="font-size: 12pt;">The target.k file contains only the target nodal coordinates. The material is an Ogden hyperelastic model, and the forward analysis is implicit.</span></p><p data-start="522" data-end="720"><span style="font-size: 12pt;">LS-DYNA generates iter2.inp and iter1.guess, but it does not generate iter1.algo. When iteration 2 starts, LS-DYNA crashes in REPNOD, apparently while attempting to read the missing iter1.algo file:</span></p><p data-start="722" data-end="780"><span style="font-size: 12pt;">forrtl: severe (157): Program Exception - access violation</span></p><p data-start="782" data-end="909"><span style="font-size: 12pt;">lsdyna_mpp_dp_msm REPNOD 37088 dynkab.F</span><br><span style="font-size: 12pt;">lsdyna_mpp_dp_msm KEYWRD 12115 dynkaa.F</span><br><span style="font-size: 12pt;">lsdyna_mpp_dp_msm LSINPUT 1482 dynm.F</span></p><p data-start="911" data-end="1027"><span style="font-size: 12pt;">I have tested both the MPP and SMP executables, and the same issue occurs, so it does not appear to be MPI-specific.</span></p><p data-start="1029" data-end="1285"><span style="font-size: 12pt;">Could anyone please advise whether iter1.algo should always be generated automatically after iteration 1? Is this a known issue with *CONTROL_REFERENCE_CONFIGURATION in the 2025 R2 Windows build?</span></p><p data-start="1287" data-end="1393"><span style="font-size: 12pt;">Any guidance would be greatly appreciated. I understand the issue with this keyword has been brought up some time ago - may I inquire about any fix since then, or any successful examples to share? I can provide the input and iteration output files, in any feasible way, if required.</span></p><p data-start="1395" data-end="1408"><span style="font-size: 12pt;">Thank you.</span></p>

I am using LS-Dyna R11.2.2. I am performing an explicit simulation and then perform the implicit springback step.  I have successfully used the 2-step approach, in which I call the *INTERFACE_SPRINGBACK_LSDYNA in the first simulation, and then include the generated dynain file in the second simulation. This approach works.  However, I have tried to combine both approaches using *INTERFACE_SPRINGBACK_SEAMLESS, and while the simulation runs, it will not output the ASCII dynain or lsda binary files. I have tried to include both *INTERFACE_SPRINGBACK_SEAMLESS and *INTERFACE_SPRINGBACK_LSDYNA in the same keyword, but then the implicit simulation does not run at all, though it outputs some dynain file.  Is there any conflict between using *INTERFACE_SPRINGBACK_SEAMLESS and generating the corresponding dynain file automatically?

Hello, I am investigating an output-mapping and energy-conservation issue in an academic civil/mining solver-verification unit test. This is not a weapon or defence model. Environment: Ansys LS-DYNA Student, Windows x64 SMP double precision R16.1-180-gd50332dbe5, one thread Units: mm-ms-g-N-MPa 2D axisymmetric S-ALE, 12,423 cells 808 cells initially filled with EMULSION *MAT_HIGH_EXPLOSIVE_BURN (MAT_008, BETA=0) with *EOS_JWL NEIPH=4, D3SALE=0, DMPMSH=1 Termination time: 0.100 ms I compared two input decks that are identical except for the lighting time LT in *INITIAL_DETONATION: v3: LT=0.0 msNormal termination, but the maximum absolute total-energy drift calculated from all original GLSTAT states is 15.995852%. E2: LT=0.200 ms, after the 0.100 ms termination timeNormal termination, and the maximum absolute total-energy drift is only 0.00049064%. No interpolation, state deletion, filling, smoothing or renormalization was used. However, I cannot yet prove from the original output that burn fraction F remained zero in all 808 explosive cells in E2. The binout contains only nine tracers, and the R16.1 manuals do not identify the exact full-field burn-fraction mapping for 2D axisymmetric S-ALE. Could an LS-DYNA specialist please clarify: With D3SALE=0, where exactly is the per-S-ALE-cell MAT_008/JWL burn fraction F stored in d3plot? Please identify the documented array/record location and the LS-PrePost Fcomp or Misc component. With D3SALE=1, what is the corresponding LSDA path and array name in d3sale? Is NEIPH=4 sufficient for full-field burn-fraction output for MAT_008 + EOS_JWL in 2D axisymmetric S-ALE? Does HISNOUT=3 provide valid labels for this field? Are these output fields identical between the R16.1 Student and commercial executables? Is the approximately 16% energy loss for LT=0, compared with approximately 0.00049% for LT=0.200 ms, a known R16.1 limitation, defect or energy-accounting behavior? Is there a recommended patch level, diagnostic output or control setting? I am not claiming that the A/B comparison uniquely proves that burn, JWL or advection is the cause. I first need the officially supported full-field burn-fraction mapping. Thank you.

My LS Prepost keeps on crashing when I hit on apply after setting  ua a boundary condition, is there any other way to do it ao maybe I am making a mistake, please help.

<p>I am working on implementing a user defined hyperelastic material which needs a push forward operation. I saw the push_forward_2s(), I'm using the scalar version for now, in Manual 1 Appendix A. However, it does not describe in sufficient detail for me to understand what this function does. </p><p>Detailed questions:</p>What stress is this function expecting as input? First Piola-Kirchhoff? Second Piola-Kirchoff? Something else? Does it account for the change in volume from the reference to the current configuration with 1/detF or do I need to handle that outside of this subroutine? What stress is the ouptut from this function? I am assuming Cauchy but I would like confirmation.<p>I have implemented my own push forward code in the user-defined material to go from the second PK stress to Cauchy stress using the well known equation sig = 1/detF * F * 2PK * F^T. However, the results do not match. It is also possible I have made a mistake in my code but the push forward is relatively straight-forward. </p><p>Thanks in advance.</p>

<p>Hello there,</p><p class="PDq2pG_selectionAnchorContainer" data-start="113" data-end="180">I have a question regarding the outputs when using LS-DYNA/MPP. I would like to obtain the displacement of the rigid body, so I requested this output in the input file. In the output, I loaded the binout0000 file and selected rbdout> rigid_body> global_dz from the main branches in the binout file (the result is shown below).</p><p></p><p data-start="444" data-end="582">However, I noticed that there is another rigid body displacement output available under matsum> SPHnode> z_rbdisplacement (shown below).</p><p></p><p data-start="587" data-end="758">Since both outputs appear to represent rigid body displacement but give different values, could you please let me know which one is the correct displacement output to use?</p><p data-start="763" data-end="773">Thank you,</p>

<p><p>I am working on implementing a user defined material model that requires access to the displacement field directly to include some atypical components like the hessian of the displacement field. </p><p>How can I pass the element nodal displacement field into the user defined material subroutine? I looked through all the documentation and a lot of the code in the object file version of LS-DYNA. However, many of the variables are not defined, e.g. in commented code sections, and the naming convention doesn't not provide a sufficient description to make an educated guess. </p><p>I know it seems a little odd but the deformation gradient is right there which is one step removed from the data I need. I would think the displacements are nearby I just don't know where to find them.</p><p>Thanks in advance for any assistance you can offer. </p></p><p> </p><p><p> Edit 7/28/2026 12:15p </p></p><p><p> After some further digging it seems like it might be possible to define my own solid element type which has a call to the usrmat() subroutine. The user defined element subroutine has access to the nodal displacement field which I could then pass to the main user defined material subroutine usrmat() and on into the individual umatxx() subroutines where the material behavior is actually defined. Seems like a long and tedious work around. </p></p><p><p> I am also wondering if the nodal displacements are more directly accessible to the user defined material subroutine through a common block data structure. That would be very straightforward and simple to use. More digging on this required. I was hoping that's how the displacement field was referenced in the user defined element subroutine and I could just copy it but alas tis not the case. </p></p><p><p> Thinking about it more here I am going to assume that the usrmat() function called from the standard solid element subroutine does not pass the nodal displacements to the usrmat() subroutine, and I would need access to that code to make that modification. Which makes me think that the user defined element is the way to go. However, if I update the subroutine call format are other parts of the code like the standard solid element function call going complain; I'm going to try this first because it is the simplest solution. If that doesn't work then, I think one could define a custom usrmat_custom() which would be called by the user defined element and could be passed the displacement field. </p> </p><p><p> Here's to chasing little white rabbits and jabberwockies. </p></p>

Hello everyone. I'm modeling a huge beam mesh model.It's a 100x50 meter mesh made of 6mm wire, using a beam model with a 100x100 mm mesh size. I created a Workbench-Share  to combine all the beams into one piece and connect all the nodes. This resulted in a gigantic beam mesh model.When I open Workbenches Dyna, it takes a very long time to load. I've waited three days, but it still hasn't loaded. How do I load it? Are there any ways to speed it up? Thanks.

<p>I just downloaded 2026R1 ansys student license and the lsdyna split input file feature seems to have stopped working?  Does anyone else also see this?</p><p>literally just did this on 2025R1 and it works. </p><p></p>